Using PIX4Dcatch to generate a Gaussian Splat model

This article explains how to generate a Gaussian Splat model by capturing images with PIX4Dcatch and uploading to PIX4Dcloud for processing.

What is Gaussian Splatting? 

Gaussian Splatting is a technique used to enhance the visual quality of 3D outputs, such as point clouds and 3D textured meshes, by producing a highly photorealistic representation of a scene.

In this approach, the scene is represented by millions of particles, known as Gaussians, each defined by attributes such as position, orientation, scale, opacity, and color.

The process begins with an initial photogrammetric workflow, which calibrates the images and generates an initial point cloud. Each point in this point cloud is converted into a Gaussian, resulting in an enhanced point cloud that ultimately leads to a more accurate and visually appealing point cloud and 3D textured mesh than traditional photogrammetric methods.

What are the requirements to generate a Gaussian Splat model?

The following conditions are needed to be able to process with Gaussian Splatting:

• Own an active PIX4Dcloud or PIX4Dcloud Advanced license
• Have enough PIX4Dcloud credits.
• Install PIX4Dcatch on a mobile device with a LiDAR sensor.
• The project must contain between 50 and 2000 images.
• Upload the images from PIX4Dcatch to PIX4Dcloud for processing.
  
  

How to select a Gaussian Splat model as output

After you have successfully captured your project, as detailed in the How to capture a project article, you will need to choose the "Gaussian Splatting" option to process your project with the Gaussian Splatting model.

Once you upload your project to PIX4Dcloud, it will be processed using the Gaussian Splatting technique.

How to see the results in PIX4Dcloud

Once the project has been processed, it can be accessed in PIX4Dcloud the same way as any other project.

If Gaussian Splatting has been selected as output in PIX4Dcatch, the hierarchy will contain a new 3D output Gaussian Splat that provides a photorealistic visualization of the scene.

In order to visualize it, enable the Gaussian Splat layer.

 

When the Gaussian Splat layer is enabled, the following message will pop up warning the user about some slowdowns in the performance that the display of the Gaussian Splat model may cause:

 

How many credits it consumes

The Gaussian Splatting process additional processing power and time. That is why a project using this technique will incur an additional flat fee of 5 credits, regardless of the number of images.

More information about the credit system here.

How to capture images to generate a Gaussian Splat model

To capture high-quality images for Gaussian Splatting (GS) in PIX4Dcatch, follow these best practices to ensure optimal results:

Plan Your Path

Object-Centric Scenes

• Move in an elliptical path with the phone always pointing towards the central object.

• If capturing multiple loops around the object, vary the height and viewing angle slightly.

• Avoid taking multiple images from nearly identical positions and angles, as they add redundancy without improving quality.

Flat Scenes

• Move in linear paths at a slanted angle (30–40°) to the surface.

• Ensure overlapping paths for complete coverage.

• Avoid capturing only perpendicular views, as this limits 3D reconstruction quality.

During Capture

• Move the phone slowly and steadily to avoid motion blur.

• Avoid sudden movements or strange motions that can distort depth calculations.

• Ensure the object fills the frame while maintaining a consistent distance.

• Use good lighting conditions; avoid shadows and moving objects that could cause inconsistencies in the reconstruction.

• If possible, include LiDAR depth maps, as they significantly improve GS results.

Save and Review

• Save the project and review the capture report to check image quality and ensure adequate scene coverage.

Following these steps will help ensure high-quality data for generating a Gaussian Splat model.

Which projects benefit from the Gaussian Splatting Technique

The Gaussian Splatting technique significantly enhances the accuracy and realism of 3D modeling, particularly when dealing with objects that have intricate geometric details and fine structural elements.

This technique proves especially effective for capturing complex shapes, which are often challenging to represent using more traditional approaches. Additionally, it excels in handling thin structures like pipes, wires, and other elongated elements that require high precision to avoid visual artifacts or loss of detail.

This technique provides a smoother, more continuous representation of object surfaces, improving visual fidelity.

Comparison between Gaussian Splat results versus non-Gaussian Splat 

This section compares several projects processed with and without Gaussian Splatting enabled in PIX4Dcatch. 

Trenches

 

Complex structures